[
https://issues.apache.org/jira/browse/HDFS-1473?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12966614#action_12966614
]
Todd Lipcon commented on HDFS-1473:
-----------------------------------
I'll fix the save() -> write() momentarily.
As for the test case, there are some tests that save in parallel, but I think
the race is really narrow. U_STR is only used very briefly before it's written
out to disk, so the likelihood of threads contending is probably pretty low.
The images we write in the test cases also aren't huge - I'm more surprised
that FB hasn't seen this yet on their large image tests. Hairong/Dmytro?
> Refactor storage management into separate classes than fsimage file
> reading/writing
> -----------------------------------------------------------------------------------
>
> Key: HDFS-1473
> URL: https://issues.apache.org/jira/browse/HDFS-1473
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Reporter: Todd Lipcon
> Assignee: Todd Lipcon
> Fix For: 0.22.0, 0.23.0
>
> Attachments: hdfs-1473-followup.txt, hdfs-1473-prelim.txt,
> hdfs-1473.txt, hdfs-1473.txt, hdfs-1473.txt
>
>
> Currently the FSImage class is responsible both for storage management (eg
> moving around files, tracking file names, the VERSION file, etc) as well as
> for the actual serialization and deserialization of the "fsimage" file within
> the storage directory.
> I'd like to refactor the loading and saving code into new classes. This will
> make testing easier and also make the major changes in HDFS-1073 easier to
> understand.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
